Abstract: Objective To construct a nomogram model for postoperative arteriovenous fistula (AVF) dysfunction in hemodialysis patients.  Methods  A total of 126 hemodialysis patients undergoing AVF operation in our hospital from July 2021 and July 2023 were taken as the research subjects. They were divided into dysfunction group (n=31) and normal function group (n=95) according to the postoperative AVF function. Their clinical data were collected. The influencing factors for postoperative AVF dysfunction in hemodialysis patients were investigated by logistic stepwise regression analysis. A nomogram model of  risk prediction for postoperative AVF dysfunction in hemodialysis patients was constructed, and its predictive value was evaluated using receiver operating characteristic (ROC) curve. Results  The incidence of AVF dysfunction in the 216 hemodialysis patients was 24.60% (31/126). The ratios of ≥60 years old, diabetes mellitus and hypotension during dialysis, and  the values of platelet count, LDL-C, blood phosphorus, and calcium and phosphorus product were higher in dysfunction group than in normal function group (t/χ2=4.946, 4.527, 10.333, 9.789, 7.164, 16.065, and 18.817 respectively, P=0.026, 0.033,＜0.001,＜0.001,＜0.001,＜0.001 and＜0.001 respectively), while serum albumin was lower in dysfunction group than in normal function group (t=7.997, P＜0.001). Multivariate logistic regression showed that age≥60 years old (OR=2.061, 95% CI: 1.431～2.967, P＜0.001), combined with diabetes mellitus (OR=2.776, 95% CI: 1.814～4.247, P＜0.001), hypotension during dialysis (OR=2.286, 95% CI:1.542～3.390,P＜0.001), higher platelet count (OR=2.499, 95% CI: 1.643～3.802, P＜0.001) and higher calcium and phosphorus product (OR=2.234, 95% CI:1.528～3.268, P＜0.001) were the independent risk factors for postoperative AVF dysfunction in hemodialysis patients. The C-index was 0.738 (95% CI: 0.687～0.789) for predicting postoperative AVF dysfunction in hemodialysis patients. ROC curve for the nomogram to predict postoperative AVF dysfunction in hemodialysis patients showed that the area under the curve (AUC) was 0.792 [95% CI：0.792 (0.741～0.843)], with the specificity of 73.05% and the sensitivity of 78.14% (Z=11.257, P＜0.001). Hosmer-Lemeshow goodness of fit test was 8.746 (P=0.394).  Conclusion  The nomogram based on age, combined diabetes, hypotension, platelet count, calcium and phosphorus product had a higher predictive value for the risk of postoperative AVF dysfunction in hemodialysis patients.

Key words: Hemodialysis, Arteriovenous fistula dysfunction, Risk factor, Nomogram model